基于内容特征的视频半脆弱水印算法

摘要

Because of the large number of watermark embedding,using semi-fragile video water-marking for content authentication demands a higher level of operating speed compared with digital image water-marking. In this paper,we improve the spatial semi-fragile watermarking algorithm which based on lumi-nance blocks,and the algorithm can keep a better robustness as well as embedding speed. By using the optimized algorithm,which is based on the feature that two adjacent luminance blocks has a constant ener-gy relationship in the non-malicious situation,the watermarking sequences are produced and embedded in-to DCT coefficient. This algorithm has its advantages,such as,does not involve specific encoding format, does not need original video for authentication,and even can localize the tampered region quickly. The simulation experimental results demonstrate the embedding speed reach 18. 37MB/s,and with invisible quality loss,meanwhile,it shows that the algorithm has good realtime response.%相比数字图像水印，在采用半脆弱水印技术进行视频内容认证时，由于嵌入水印量比较大，对算法的运行速度提出了更高的要求。本文改进了基于亮度分块的数字视频空域半脆弱水印算法，能保持水印的鲁棒性不变，且提高了水印的嵌入速度。基于相邻亮度分块的平均值大小关系在非恶意操作前后保持不变的特点，本文运用改进算法生成水印序列，并在时域上将水印嵌入DCT系数中。该算法具有不涉及具体编码格式，无需原视频参与认证，快速定位篡改区域的优点。仿真实验结果表明水印嵌入达到了18.37MB/s的速度，对画质影响不可见，实验结果同时证明了算法具有良好的实时性。